J . Am. Chem. Soc. 1992, 114, 10941-10950
10941
Electronic State-Selected Reactivity of Transition Metal Ions: Co+ and Fe+ with Propane Petra A. M. van Koppen,* P. R. Kemper, and Michael T. Bowers* Contribution from the Department of Chemistry, University of California, Santa Barbara, California 93106. Received July 13, 1992
Abstract: Recently, we developed a "chromatographic" technique to determine the electronic state distribution for transition metal ions. This method allows the study of state-selected reactions. In this paper we report the quantitative determination of rate constants and branching ratios for state-selected Co+ and Fe+ reacting with propane. Observed rates for adduct formation as well as H2 and CHI elimination channels were strongly dependent on the electronic configuration of the metal ion. Co+ ions were formed by electron impact on either Co(CO),NO or C O C ~ ( C Oor) ~by surface ionization of CoCI2. The Co+ electronic state population is a function of both the electron energy and the precursor used, and can be varied from 36% ground state to 97% ground state. Thus, we can measure reaction rate constants over a wide range of ground- and excited-state populations and extrapolate to 100%ground- or excited-state CO+to obtain the state-specific reaction rates. Under our experimental conditions ( Torr of C3Hsin 1.75 Torr of He), adduct formation is the dominant product for the a3F 3d8 ground state of Co+, with only small amounts of elimination products observed. The 4s3d7excited states (a5Fand b3F) show greatly reduced clustering (due to the repulsive 4s electron) and enhanced elimination channels. Fe+ was formed by electron impact on Fe(CO),. Again the electronic state population was varied by varying the electron energy. Absolute rate constants were obtained for the 6D ground state as well as for the 4F and 4Dexcited states of Fe+ reacting with C3H8. Adduct formation is the dominant product for the 6D4s3d6 ground state of Fe+ despite the repulsive 4s electron. This is due to a crossing from the ground-state surface to the Fe+(4F3d7)C3Hsfirst excited-state surface where the adduct is more strongly bound. The Fe+ 4D 4s3d6 second excited state reacts similarly to the Co+ a5F and b3F 4s3d7 excited states.
Introduction Probing the kinetics, dynamics, and thermochemistry of organo-transition-metal reactions using mass spectrometric techniques has shown these reactions to be complex1**and usually extremely dependent on the metal ion electronic ~ t a t e . ~ - It I ~is therefore important to carry out state-selected reactivity studies. Radiative lifetimes of excited-state metal ions are long (on the order of seconds) owing to parity forbidden transitions.I3 Consequently, once the excited electronic states are produced, their reactivity can be studied provided the state can be selected. The electronic state configurations and corresponding energies for Co+ and Fe+ are summarized in Table I.I4 Atomic transition metal ions produced by electron impact,j surface i0nization,6.'~~~ or laser (1) For a recent review, see: Eller, K.; Schwarz, H. Chem. Reu. 1991,91, 1121 and references therein. (2) Gus Phuse Inorganic Chemislry; Russel, D. H., Ed.; Plenum Press: New York, 1989. (3) (a) Kemper, P. R.; Bowers, M. T. J . Phys. Chem. 1991,95, 5134. (b) Kemper, P. R.; Bowers, M. T. J. Am. Chem. SOC.1990, 112, 3231. (4) Weisshaar, J. C. In Aduances in Chemical Physics; Ng, C., Ed.; Wiley-Interscience: New York, 1992; Vol. 81 and references therein. ( 5 ) Sanders, L.; Hanton, S.; Weisshaar, J. C. J . Phys. Chem. 1987, 91, 5145. Sanders, L.; Hanton, S. D.; Weisshaar, J. C. J . Chem. Phys. 1990,92, 3498. (6) Armentrout, P. B. Annu. Reu. Phys. Chem. 1990, 41, 313 and refer-
ences therein. (7) Armentrout, P. B. In Gus Phuse Inorgunic Chemisrry; Russell, D. H., Ed.; Plenum Press: New York, 1989. (8) Freas, R. B.; Ridge, D. P. J. Am. Chem. Soc. 1980,102,7129. Reents, W.D., Jr.; Strobel, F.; Freas, R. B., 111; Ridge, D. P. J . Phys. Chem. 1985, 89, 5666. (9) Halle, L. F.; Armentrout, P. B.; Beauchamp, J. L. J . Am. Chem. SOC. 1981, 103, 962. (IO) Elkind, J. L.; Armentrout, P. B. J . Phys. Chem. 1986, 90, 5736. ( 1 1 ) Elkind, J. L.; Armentrout, P. B. J . Phys. Chem. 1987, 91, 2037. (12) (a) Sunderlin, L.S.; Aristov, N.; Armentrout, P. B. J. Am. Chem. SOC.1987, 109, 78. (b) Aristov, N.; Armentrout, P. B. J. Am. Chem. Soc. 1986, 108, 1806. (13) Garstang, R. H. Monr. Nor. R. Astron. SOC.1962, 124, 321. (14) (a) Moore, C. E. Atomic Energy feuels; U S . National Bureau of Standards: Washington, DC, 1952; Circ. 467 (US.National Bureau of Standards). (b) Sugar, J.; Corliss, C. J . Phys. Chem. ReJ Dura 1981, I O , 197, 1097. (c) Sugar, J.; Corliss, C. J . Phys. Chem. ReJ Dura 1982, 11, 135. (15) Sunderlin, L. S.; Armentrout, P. B. J . Phys. Chem. 1988, 92, 1209. (16) Schultz, R. H.; Elkind, J. L.; Armentrout, P. B. J . Am. Chem. SOC. 1988, 110, 411.
0002-7863192115 14-10941$03.00/0
Table I. Electronic States of Fe+ and Co+ ion state configuration
Fe+
a6D a4F
4s3d6
a4D
4s3d6 3d7
a4P a 2G
co+
a3F
aSF b3F a'D a3P Reference 14; averaged over J
3d7 3d7 3d8 4s3d7
4s3d7 3d8 3d8
energy'' (eV) 0.052 0.300 1.032 1.688 1.993 0.086 0.5 15 1.298 1.445 1.655
levels.
v a p o r a t i ~ n are ~ ~ ~formed '~ in a mixture of ground and excited states. With electron impact (EI) the excited electronic state population is a strong function of the electron energy, increasing rapidly with increasing electron energy up to 40-50 eV. Laser vaporization also produces a significant amount of electronically excited ions. With surface ionization (SI), a Boltzmann distribution of ground and excited states is produced. Only with resonant multiphoton ionization (REMPI)4.S,19 can pure groundand excited-state metal ions be produced. REMPI is limited, however, owing to the complexity and difficulty associated with this technique. Recently, we developed a simple way to characterize populations of electronically excited metal ions using a 'chromatographic" t e c h n i q ~ e . ~ The , ~ ~ valence electron configurations for atomic transition metal ions, which are 3d" or 4s3dn-I, exhibit large differences in mobility. These differences in mobility give rise to a spatial and temporal spread of the metal ions in different electronic states as they diffuse through a buffer gas. If a mass-selected ion beam is pulsed into the reaction cell containing ~
(17) Loh, S . K.; Fisher, E. R.; Lian, Li; Schultz, R. H.; Armentrout, P. B. J. Phys. Chem. 1989, 93, 3159. (18) Cody, R. B.; Burnier, R. C.; Reents, W. D., Jr.; Carlin, T. J.; McCrery, D. A.; Lengel, R. K.; Freiser, B. S. Int. J. Moss Spectrom. Ion Phys. 1980, 33, 37. (19) Hanton, S . D.; Noll, R. J.; Weisshaar, J. C. J . Phys. Chem. 1990, 94, 5655. (20) Kemper, P. R.; Bowers, M. T.J . Am. SOC.Muss Specrrom. 1990, I , 197.
0 1992 American Chemical Society
10942 J. Am. Chem. SOC.,Vol. 114, No. 27, 1992
van Koppen et al.
a buffer gas, ions with different electronic configurations separate as they diffuse through the cell and are observed at different times in the arrival time spectrum. The excited- and ground-state populations can be determined by integrating the corresponding peak areas. The electronic state population is a function of both the electron energy and the neutral precursor used and can be varied over a wide range. Reaction rate constants are measured as a function of the percent ground and excited state to obtain the state-selected rate constants for reaction (by extrapolation to 100% ground- and excited-state M+).2' In this paper we will use this technique to determine the reactivity of ground and excited states of Co+ and Fe+ with propane. We report the quantitative determination of rate constants for all the observed products which include adduct formation as well as H2, CHI, and C2H5elimination channels. The implications of these results on the potential energy surfaces will be discussed.
(&+I* 4dd7 asF. b3F
Co+ 3ds a3F
Experimental Section Details of the instrument used in these experiments have been described20 and are only briefly outlined here. The instrument used coupled
a reverse-geometry mass spectrometry operating at 5 kV to a highpressure, temperature-variable drift cell that operates at thermal energies. The first-stage mass spectrometer is a home-built instrument with the same dimensions and ion optics as a V.G. Instruments ZAB-2F. The ion source is a standard V.G. Instruments electron impact ionization source. The source was modified to do surface ionization, with a design similar to that of Armentrout and co-workers.12b Co+ and Fe+ ions are formed ~ , Fe(CO),, reby electron impact on C O ( C O ) ~ N Oand C O C ~ ( C O )or spectively. In addition, Cot ions are formed by surface ionization of CoCI,. In the case of Fe+, the largest isotope is m/e = 56 (91.8%) which conicides with the mass of (CO),+ or Fe(C0)22+,at least one of which is also formed by electron impact on Fe(C0)5.22 As a result, we used the next largest Fe+ isotope, m / e = 54 (5.8%) in all the experiments. Upon exiting the source, the ions are accelerated to 5 kV and mass selected using the double-focusing reverse geometry mass spectrometer. The ions are then decelerated to 2-3 eV kinetic energy and are focused into the high-pressure drift cell containing 1.75 Torr of helium buffer gas. The ions are quickly thermalized by collisions with He and drift through the reaction cell at constant velocity because of the presence of a uniform drift field. Ions react with a trace (- 1 X lo-, Torr) of C3H8present in the He, exit the cell, and are quadrupole mass analyzed and detected. Reaction times typically range from 200 to 600 ps corresponding to drift fields of E / N I3.5 X lo-'' Vcm2. These fields do not perturb the ion translational temperature more than a few degrees.20 The electronic state populations are largely obtained from the ion arrival time distribution (ATD). The ATD for Co+ or Fe+ is measured by pulsing the mass-selected ion beam into the drift cell (pulse width 1-3 ~s). The pulse simultaneously triggers a time-to-pulse-height converter ramp. Ions that exit the cell are collected as a function of time, giving the arrival time distribution. Ions that have different mobilities have different drift times through the cell and appear as different peaks in the ATD. With Co+ no excited-state deactivation occurred and the integrated ATD peak areas equal the populations of the electronic configurations. In the Fet case, deactivation does occur, and a fitting of theoretical and experimental ATDs was required to determine the electronic configuration populations (described below). In both the Fe+ and Co+ experiments, two states with 4s3d"' configurations were present, and additional information from the reaction was required to determine the state populations. This point is discussed below. Additional information regarding individual electronic state reactivities is obtained from product ion ATDs. In these experiments the Co+ or Fet is pulsed into the cell, but a particular product ion is collected instead of the bare metal ion. These product ATDs have separate components corresponding to the reaction of the 3d" or 4s3d"'l configurations to form the particular product. This provides a separate, semiquantitative determination of the relative efficiencies with which the M+ states react to form the product ion. The use of this technique in the Co+ experiments is discussed below. The state-specific rate constants are determined by studying the reaction as a function of the state populations. These populations are
-
(21) van Koppen, P. A. M.;Kemper, P. R.; Bowers, M.T. J . Am. Chem. Soc. 1992, 114, 1083.
(22) We observe a doublet in the arrival time distribution for s6Fe+. The longer time peak in the doublet is not present for "Fe+. Both short time peaks (for 56Fe+and s4Fe+)show tailing to longer times, suggesting deactivation, a point discussed in detail later in this paper. High resolution mass spectra taken by Oriedo and Russell indicate the second peak, (nominal mass m / e = 56) is due to the Fe(CO)?+ impurity (private communication).
105
161
217
Time (ps) Figure 1. Cot arrival time distribution (ATD), 300 K, 50-eV electron energy, Co+ from Co(CO),NO. The absence of Co+ arrival times intermediate between those of the excited and ground state indicates deactivation does not occur from the excited state to the ground state. The integrated peak areas equal the populations of the electronic state configurations. altered by changing both the electron energy and the precursor used. For example, at 50 eV, electron impact on Co(CO),NO produces 36% ground-state Co+, whereas C O C ~ ( C Oproduces )~ 83% ground-state Co+. With the electronic state population determined, [M+]/[M+], is measured as a function of time to obtain the total rate constant, k,,,. We measure k,,, as a function of percent ground-state M+ and extrapolate to 100% ground-state and 100%excited-state M+ to determine the relative rates of reaction. Product distributions are measured as a function of percent ground-state M+ to obtain individual rate constants. The accuracy of the absolute total rate coefficient measurements is estimated to be within *30%.20 T H e relative rate coefficient measurements, however, are much more accurate (&lo%). The pressure of propane was varied from 1 to 4 X lo-, Torr, yielding rate coefficients unchanged to within *15%.
Results and Discussion I. ArrivalTiiDistributiom: Co++C&
a. Electroaicstate
Populations. A typical ATD for Co+ is shown in Figure 1. Two
peaks are observed corresponding to ground (3d8) and excited (4s3d7) electronic state configurations. The excited-state Co+ contains a 4s electron which is larger and more repulsive than the ground state which contains only 3d electron^.^ The reduced attraction to He gives the excited-state Co+ a greater mobility than the ground state, causing the excited state to appear earlier in the ATD than the ground state. The two peaks are baseline resolved. The absence of Co+ arrival times intermediate between those of the excited and ground state indicates deactivation does not occur from the excited state to the ground state, while the ion traverses the reaction cell. Since deactivation does not occur, the integrated ATD peak areas equal the populations of the electronic configurations. In a recent study3 of first row transition metal ions, IV-XII, collisional deactivation was observed only for Fe+ and Mn+in He. These are the only cases where an excited state of 3d" configuration was observed to deactivate to a 4s3d"I ground state. Ti+ also has a 4s3dn-' ground- and a 3dn excited-state configuration and has been observed to deactivate by Tonkyn and Weisshaar.23 The ATD resolution in our previous Ti+ experiments was insufficient to observe this. In no case did excited 4s3d"l states collisionally deactivate a t thermal energies. This finding is nicely explained by a model proposed by Loh et aI.l7 Consider the potential curves for a M+-He collision when the M+ ground state ion has a 4s3d"I (23) Tonkyn, R.; Weisshaar, J. C. J . Phys. Chem. 1986, 90, 2305.
Reactions of Co+ and Fe+ with CJHR configuration and the excited state has a 3d" configuration. The 4s3dn-Ipotential energy curve will become repulsive at larger internuclear distance than the 3d" excited configuration, owing to the repulsion between the 4s electron and the filled Is2 shell of He. Thus the potential energy curves for the two states in the collision will cross, perhaps at a low collision energy, providing a means for deactivation in the collision. For 3d" ground states and 4s3d" excited states, however, the curves will not cross until much higher collision energies, and consequently, excited 4s3dn-' configurations should not easily deactivate, as we observe experimentally. On this basis, the 3ds a l D and a3P states of Co+, which lie only 0.147 eV and 0.357 eV above the 4s3d7 b3F state, respectively, should rapidly collisionally deactivate to the b3F state. Since such deactivation is not observed in the ATD for Co', and a3P and higher lying excited states are assumed not to be present (52%) under the high-pressure conditions of our experiment. Consequently, the peak corresponding to the 3d8 electronic configuration contains only the ground state. The excited state, however, may be a composite of two states, the a'F and the b3F, both of which have 4s3d7 electron configurations. Translational energy spectroscopy experiment^^^ have confirmed the presence of the b3F second excited state in Co+ when formed by electron impact on C O ( C O ) ~ N O .No ~ ~information was obtained regarding the a5F state.26 To obtain information regarding the reactivity of the a5Fstate with propane, Co+ was formed by surface ionization of CoCl2. Unlike electron impact, surface ionization of CoCl, on a resistively heated rhenium ribbon at approximately 2300 K does not have enough energy to form measurable amounts of the b3F second excited state. The electronic state population obtained by integrating the ATD peak areas is 15 f 1% a5F first excited state and 85 f 1% a3Fground state in agreement with the calculated Boltzmann distribution at this temperature. The a S Fstate was found to react with propane at a rate very similar to that of the combined b3F and aSF states formed by electron impact (the product ATDs for all channels were the same for Co+ formed by electron impact and surface ionization; see following section). Consequently, the presence of any a'F Co+ formed by electron impact will have a negligible effect on our reported Co+(b3F)+ C3Hs rate constants, although the reported fraction of b3F would be in error. b. Product Ion Peak Shapes. Product ATDs for the CoC3Hs+ adduct are shown in Figure 2. Figure 2a shows the ATD for CoC3H8+ions formed in the ion source. The adduct is mass selected and injected into the reaction cell containing only helium. In this case the ATD consists of only one peak. This is due to the fact that the Co+C3Hsadduct is much larger than the Co+ atomic ion, and the resultant differences in the mobility due to different electronic state configurations of Co+ are negligible. Because of its larger size, the adduct has a lower mobility and shows up at longer times in the ATD (Figure 2a) than either (Co')' or Co+ (Figure 2c). If Co+ is injected into the reaction cell, where it reacts with propane to form the adduct, the resulting ATD for the CoC3Hs+product ion is broad and has two components as shown in Figure 2b. The arrival time of the adduct is a function of the position at which the reaction took place in the cell. The onset of the (Co+)*C3H8and CO+C3H8 ATDs correlate exactly to the ATDs of (Co+)* and Co+ in Figure 2c, as indicated by the dashed arrows. We interpret this result as follows. If Co' reacts with propane at the end of the reaction cell, the arrival time for CoC3Hs+formed will correspond to the arrival time of Co+. The longest arrival times correspond to the ATD of the adduct formed at the entrance of the cell. This corresponds to the ATD of Co+C3H8formed in the ion source (Figure 2a) as shown by the dashed arrow. Ions which react (24) Illies, A. J.; Bowers. M. T. Chem. Phys. 1982. 65, 281. (25) Hanratty, M. A.; Beauchamp, J. L.; Illies. A. J.; van Koppen. P. A. M.; Bowers, M. T. J . Am. Chem. Soc. 1!388, 110. 1. (26) The a'F state lies only 0.43 eV a b v e the a'F state (Table I). As a result. it is difficult to resolve these two states using translational energy spectroscopy. With increased resolution, the asF state may still not be ob-
served since the a5F
-
a'F transition is spin forbidden.
J . Am. Chem. SOC.,Vol. 114, No. 27. 1992 10943
A
I I
I
t 1 I 1
I
I
CO'
105
161
217
Time (CIS) Figure 2. (a) ATD for Co+C3HEand (Co+)*CJH, formed in the ion source. The adduct is mass selected and injected into the reaction cell. A single peak is observed since (Co')*C,H, and Co+C3HEhave the same mobility (see text). (b) ATD for Co+C,Hs and (Co+)*C3HEformed in the reaction cell. In this case, the Co' ion beam, containing 48% ground-state Co' and 52% excited-state (Co')', is injected into the reaction cell where it reacts with propane to form the adduct. The amval time of the adduct is a function of the position at which the reaction took place in the cell. If Co' reacts with propane at the end of the reaction cell, the arrival time for CoC3HE+formed will correspond to the arrival time of Co+. As a result, the onset of the (Co+)*C,Hs and Co+C,HE ATDs correlate exactly to the ATDs of (Co+)* and Co', as indicated by the dashed arrows. The longest arrival times correspond to the ATD of the adduct formed at the entrance of the cell. This corresponds to the ATD of C O + C ~ Hformed , in the ion source (a) as shown by the dashed arrow. The shaded area corresponds to excited state (Co')' reaction.
throughout the cell have intermediate arrival times, giving rise to the broad flat-topped peak we observe. The shaded area corresponds to excited-state (Co')' reaction. The shapes of the product ATDs are complicated. However, the calculated ATDs for the (Co+)*C3Hsand Co+C3Hs,from known transport prop erties, are expected to be broad and to decrease in intensity at long times. Based on this as well as experimentally observed peak shapes where a single state dominates the reaction, we approximate the peak shape indicated by the shaded area for the (Co+)*C3Hs adduct. Since we have roughly equal amounts of excited- and ground-state Co+ to start with (48% ground state and 52% excited state), it appears from the relative areas in Figure 2b that ground-state Co+ forms the adduct far more efficiently than does the excited state. As we increase the percentage of ground-state Co+ from 48% to 93%. the (Co+)*C3Hsadduct contribution to the ATD essentially disappears. The product ATD for CoC3H6+(H2 loss), shown in Figure 3, indicates that this product is produced much more efficiently from the excited state than from the ground state. Starting with 48%
J . Am. Chem. SOC.,Vol. 114, No. 27, 1992
-
CO' + C ~ H R
-
CO*+ C ~ H A
C O ' C ~ H+~ H2
4Rrr Ground Staw Co+
van Koppen
et
al.
CO*C2Hd + CH4
t
133
189
24 5
30 1
T i m e (ps) bl 93C: Ground State Co*
93-r Ground State Co+
t
133
189
245
30 1
T i m e (ps)
Time (ps)
Figure 3. ATDs for Co+C3H6and (Co+)*C,H, (H, loss from the cobalt-propane adduct) formed in the reaction cell. ATDs for (a) 48% and (b) 93% ground-state Co+ are shown. The shaded area corresponds to ground-state Co+ reaction. From the relative areas, excited state Co+ is shown to eliminate H, more efficiently than ground state.
Figure 4. ATDs for Co+C2H, and (Co+)*C2H, (CH, loss from the cobalt-propane adduct) formed in the reaction cell. ATDs for (a) 48% and (b) 93% ground-state Co' are shown. The shaded area corresponds to ground-state Co+ reaction. From the relative areas, excited state Co+ is shown to eliminate CH4 more efficiently than ground state.
Scheme I
reactionrates giving rise to the sum of two exponentials shown in eq 2: c o + = f(Co+)ge-k,'
ground-state Co+ (Figure 3a), the observed contribution to the CoC3H6+ATD from the ground state (the shaded area) is roughly 10%. Starting with 93% ground-state Co+ (Figure 3b), a significant increase in the ground-state contribution to the CoC3H6+ ATD is observed. These results indicate the (Co+)* excited state(s) is (are) about five times more efficient at producing Co+C3H6than is the ground state. The product ATD for CoC2H4+ (CH4 loss), Figure 4, is very similar to that of CoC3H6+ (H, loss). Again the excited state is much more efficient in comparison to ground state in eliminating CH4. These experiments give a semiquantitative determination of the relative Co+/(Co+)* efficiencies in forming a given product ion. This result is completely independent of (and complementary to) our kinetic results discussed below. n. Kinetics: Co+ + C3Ha. The mechanism assumed for Co+ reacting with C3H8involves the formation of an internally excited (Co+C,H8)*complex which can be stabilized by collisions with helium, dissociate back to reactants, or eliminate H, or CH4 (Scheme I). For one electronic state of Co+, the fractional decrease CO+/(CO+)~ is a simple exponential decay:
ktot
Co+ = (Co+)#-kimr kfk,(He) + k'kr = kb + k,(He)+ k ,(C3H8)
+ (1 -J)(co+)#-kea'
(2)
In this expression, Co+ corresponds to the sum of ground- and excited-state Co+,fis the fraction of ground-state Co+, and k, and k,, are the ground- and excited-state rate constants, respectively. In the low conversion limit, used in our experiments, the exponential decay of both ground- and excited-state Co+ is well described by a linear function (i.e., e-krz 1 - kt for kt 2.5 0.21 0.26 k~c+/k(~c+)** 40 'The accuracy of the absolute rate coefficient measurements is estimated to be. within 30%F0 The relative rate coefficient measurements, however, are much more accurate (f10%)F9 He stabilization is important because of the low propane pressure. CRateconstants in units of 1O-Il cm3/s. dDeactivation rate constant with helium buffer gas, in units of cm3/s. 'The Langevin rate constant, kL = 1.19 X lo4 cm3/s.
be superimposed on the 6D/4D peak. However, as noted above, estimates of these state populations under multicollision condit i o n ~range ~ ~ from -0% to 5%. If a relatively small percentage of these highly excited 4s3d6 states are present, the consequences are as follows. First, the large energy available (2.59-3.72 eV) might open other reaction pathways; however, no products were found other than those known to come from the 6D, 4F, and 4D states. Second, without a low-lying Fe+ 3d7 state available for curve crossing, we do not expect these highly excited 4s3d6 states to cluster with C3H8,by analogy with the excited Co+ 4s3d7results. Thus, the elimination reactions appear to be the only pathways available to any Fe+ states above the 4D. From Oriedo and Russell's data it appears that if any highly excited Fe+ states are present, they are of 4s3d6 configuration. From the above discussion, these states react similarly to the 4D state if they react at all. As discussed for the 4s3d7 states of Co+, mixing with the 3d7 surface allows for C-H or C-C bond activation to occur. Because of the large energy difference between the highly excited 4s3d6 states an the 3d7 states of Fe+, state mixing is not probable. Elimination reactions may thus be very inefficient. This validates our analysis except to note that part ( 100 ps.
-
of elimination products increases to 10% of the total products. Also, the C2H5elimination channel is now present. As discussed above in section IV, we assume the 4D state has a clustering rate about 2.5% that of ground-state Fe+, by analogy with the excited 4s3d7 Co+ states. In fact, the excited Fe+ 4s3d6 states may cluster somewhat less efficiently owing to the larger Fe+ ion size.35 In any case, we expect the 4D clustering rate constant to be small, and in this case comparison of the known Fe+ ground-state elimination rate constants with those observed at 50 eV (when both 6D ground state and 4D states are present) allows a unique determination of the 4D population and rate coefficients. These are listed in Table 111. c. Fe' 4F 3d7 First Excited State. Analysis of the 4 F state begins with the initial state populations and the collisional deactivation rate. As discussed, by theoretically fitting the experimental Fe+ ATDs, the collisional deactivation rate constant was determined to be 9.5 X lo-" cm3/s and the initial fraction of 4F was determined to be 22.5% at 50-eV electron energy. The procedure used to extract the elimination and adduct formation rate coefficients involved calculating theoretically the observed (35) Barnes, L. A,; Rosi, M.; Bauschlicher, C. W., Jr. J . Chem. Phys. 1990,93, 609.
Reactions of Co+ and Fe+ with C3H8 product distribution for Fe+ (all states) reacting with C3HB as a function of time. This requires an exact solution of the kinetic equations, details of which are given in the Appendix. Since the only unknowns are the reaction rate coefficients, matching the experimental and calculated product distributions at long times (no 4 F present) and at short times (,F present) determined the 4F rate constants. This fitting, however, is insensitive to the adduct formation rate constant since the total rate constant (which is the sum of adduct formation, elimination, and deactivation rate constants) is approximately equal to the deactivation rate constant (within 1%; see Appendix, section 2). Thus,the adduct formation rate constant of the state must be obtained in an alternate way. An upper limit is derived from the Fe+.C3Ha product ATD. This ATD shows no observable adduct derived from a 3d7 electronic state configuration of Fe+ (e.g., Fe+ 4 F 3d7). From this we estimate the 4Fcontributes less than 4% to the clustering channel. Together with the known initial population, this sets a limit of k3(He) < 20 X 10-” C”/s. The rate constants are summarized in Table 111. VI. Discussion of Reaction Rates: Fe+ + C&. Reaction rate constants for the 6D, 4F, and states of Fe+ with C3Hs at 300 K and 1.75 Torr of He are summarized in Table 111. The major product under these conditions is adduct formation (96% for ground state Fe+) being formed at 40% of the collision rate. These results are in good agreement with Tonkyn et where the adduct was also found to be the major product (94% at 300 K and 0.75 Torr of He) being formed a t 53% of the collision rate. The rapid collisional deactivation of the Fe+ state to the ground state, indicates that a curve crossing occurs in the Fe+/He collisions. Our experimental data for Fe+ reacting with propane is also consistent with a surface crossing of the 4F state to the ground state, since relatively efficient ground-state adduct formation is found. However, the rate of adduct formation for ground-state Fe+ reacting with C3H8 is a factor of 2 smaller than that of Co+ ground state. This is not unexpected given the effectively reduced well depth of the Fe+C3Hacomplex relative to the Fe+ ground-state asymptote since the bottom of the well diabatically correlates to the Fe+ asymptote that lies 6 kcal/mol above the 6D asymptote (see Figure 9). The surface crossing implies a common C-H and/or C-C bond activation transition state, X’,for the and 6D ground state Fe+ reacting with propane (see Figure 9). Since we observe H2 and CH4 loss from ground-state Fe+, X’ must lie below the ground-state Fe+/C3Ha reactant energy. In a separate study, X* was determined to be located only 0.07 eV below the asymptotic energy of the ground-state reactants.36 The state lies 0.25 eV above ground-state Fe+ and is therefore 0.32 eV above the transition state X’. Thus,when the Fe+ state reacts with C3H8 we expect (1) deactivation to Fe+ ground state, (2) a large enhancement of the elimination channels, and (3) a decrease in the rate of adduct formation relative to that of the ground state due to competition with elimination channels. These expectations were all observed experimentally. The rate constant for adduct formation for the 4F excited state is
